Author: Joel Topf

Hi Dr. Topf,

I hope that you are doing well. My name is XXXX and I am an M2 at OUWB. I was just reviewing your Disorders of Sodium and Water Metabolism lecture again, and I have a quick question.

In a patient with decompensated heart failure presenting with hypervolemic hyponatremia, are loop diuretics appropriate here? While they address volume overload, is there a risk of worsening the hyponatremia? I know in the lecture you mentioned treating the underlying issue and volume restricting.

Great question.

The key to fixing the hyponatremia in these patients is fixing the heart failure.

Remember why the patient with decompensated heart failure is hyponatremic in the first place. The problem is not that they don’t have enough volume. They have too much volume. But because the heart is failing, the arterial circulation is underfilled. The kidneys interpret this low effective circulating volume as a threat and respond by retaining sodium and water. Most importantly for the hyponatremia, ADH is released despite the low serum osmolality.

So the patient retains water and becomes hyponatremic.

This creates what seems like a paradox: the patient is overloaded with fluid, but the kidney is acting like the patient is volume depleted.

The way out of this is to successfully treat the heart failure and restore effective circulation. And in a congested patient, that means getting rid of volume with loop diuretics.

But XXXX has identified an important wrinkle: loop diuretics change the composition of the urine.

Without a loop diuretic, a patient with heart failure and lots of ADH may make a small volume of very concentrated urine. A loop diuretic increases urine volume, but it also increases the amount of sodium and potassium in that urine. That makes each liter of urine less effective at getting rid of electrolyte-free water.

This is the concept of electrolyte-free water clearance.

Very roughly, what matters for the serum sodium is not simply:

How much urine did you make?

but:

How much water did you excrete relative to sodium + potassium?

So loop diuretics push things in two different directions. They increase urine volume, which tends to increase water loss and raise the serum sodium. But they also increase urinary electrolyte excretion, which means each liter of urine carries away less electrolyte-free water.

In most patients with decompensated heart failure, the net effect of successful diuresis is still favorable: they lose more water than electrolytes and the serum sodium rises.

But not always.

You absolutely can have a patient whose serum sodium falls during aggressive diuresis. This is particularly likely if the heart failure remains decompensated, ADH remains high, the patient continues drinking water, and the diuretic-induced urine is relatively rich in sodium and potassium.

But that possibility doesn’t mean you should avoid loop diuretics.

The main thrust of treatment is still to get the patient from decompensated heart failure to compensated heart failure. If they are congested, that requires volume removal.

And when you succeed, effective circulation improves. The non-osmotic stimulus for ADH goes away. ADH falls. The collecting duct becomes impermeable to water and the kidney starts dumping dilute urine.

Aquaresis.

At that point the serum sodium will correct.

So yes: use the loop diuretic.

Don’t let the hyponatremia scare you away from treating the disease that is causing the hyponatremia. Follow the serum sodium while you diurese, restrict water as appropriate, and get the heart recompensated.

Fix the heart failure → remove the stimulus for ADH → excrete the excess water → fix the hyponatremia.

Hi Dr. Topf,

I hope you’re doing well! I was reviewing S11: Sodium Metabolism, and I think I’m having some trouble understanding why euvolemic hyponatremia is non-edematous while hypervolemic hyponatremia is edematous. I was wondering if you could please help me correct my thought process.

Voume status is a total body sodium issue (more sodium, more edema)

Hypo- and hypernatremia is a total body water issue (more water more hyponatremia)

For euvolemic hyponatremia, such as SIADH, my understanding is that excess ADH causes water retention, which lowers serum sodium and plasma osmolality. The resulting osmotic gradient causes some of the excess water to shift from the ECF into the ICF. I was thinking that because the retained water is distributed between both compartments in the same 2/3:1/3 ratio, there isn’t a significant enough increase in ECF volume to produce edema?

Nope, these patients do not get edema and are clinically euvolemic, though if you put them on  scale they will be heavier, it is just that the extra volume is not clinically evident.

Where I become confused is with hypervolemic hyponatremia. I understand the example of heart failure: the hearts not pumping well, our organs aren’t getting perfused and so to increase blood volume we activate RAAS and ADH, causing retention of both sodium and water. However, if the serum osmolality is also low because this is a hyponatremia, shouldn’t there still be an osmotic gradient causing some of the retained water to move into the ICF, just as it does in SIADH? What makes the ECF expand enough in hypervolemic hyponatremia to produce edema?

With the increase in total body sodium all of that volume goes to the extracellular compartment, since sodium is an extracellular ion

However the retained water from the ADH does distribute across both the intracellular and extracellular compartment in a 2/3rds : 1/3rd ratio.

The ECF expands due to the RAAS induced retention of Na

I’m also confused by the statement that hypervolemic hyponatremia involves an increase in total body sodium. If decreased renal perfusion activates RAAS, and theres increased sodium reabsorption rather than sodium loss in the urine, don’t we see a net increase in sodium. So if there’s more sodium reabsorption I think I don’t get how this is a hyponatremia or is it considered a hyponatremia because even though there’s a net increase in sodium is the water increase out pacing it where the overall concentration is is low making it a hyponatremia?

There is increased retention of Na, that is what causes the hypervolemia, and there is a net retention of water, that is what causes the hyponatremia. Yes, the increase in water retention outpaces the sodium retention.

I feel like I’m missing something or thinking through this wrong somewhere so any help is super appreciated!

I think you get it better than you thought.

Thank you so much for your time and help!

Best Regards,
XXXXX

Occam’s razor: when evaluating multiple competing theories, the simpler explanation of an entity is to be preferred.

Solid logic, but not infallible.

This week I took care of a patient with a serum calcium of 17.5 mg/dL. Calciums that high are almost always malignant in nature. But we did the internal medicine thing and ran the algorithm. First branch point, “Is this PTH dependent or independent hypercalcemia?” The PTH came back at 220 pg/mL. Primary hyperparathyroidism. Case closed…or not.

A calcium of 17.5 mg/dL is extraordinarily high for primary hyperparathyroidism. In fact it would have been the highest calcium I had ever seen in primary hyperparathyroidism. I learned as a fellow, that whenever you think “this is the worst case of X I have ever seen, consider that maybe it is not the worst case of X but rather a more pedestrian case of Y.” Maybe, instead of the worst case of primary hyperparathyroidism this was two diseases instead of one.

Hickam’s dictum: Patients can have as many diseases as they damn well please.

The workup revealed a kappa/lambda ratio of 250 to 1. The patient didn’t just have primary hyperparathyroidism. They had primary hyperparathyroidism and multiple myeloma (or MGRS, the bone marrow biopsy is still pending).

The myeloma amplified the hyperparathyroidism beyond reason.

When you see the most dramatic presentation of a common codition, don’t just marvel at its severity, but ask yourself could this be somehting else entirely?

Sometimes Occam loses and Hickam wins.

Note: This post first ran on Roon. Roon is a physician only (and US only, for now) social media site. It is growing and you should check it out.

Every nephrology fellowship has the same challenge.

Every July 1, fresh fellows walk through the door. They’re smart, motivated, and terrified of upcoming responsibilities. And every program responds the same way; we compress a year’s worth of practical nephrology into the first few weeks of didactics.

We teach dialysis. Acid-base. Electrolytes. AKI. Consult etiquette. How to survive call.

Every program builds its own version of this curriculum.

A few years ago we had a simple idea: why is every fellowship reinventing the wheel? What if we pooled our resources and built something together?

That was the beginning of Nephrology Bootcamp (Link to 2024 | 2025 | 2026).

The original goal was pretty modest. Invite fellows from around southeast Michigan. Recruit some great teachers. Share the educational materials we developed independently. If one program had the best electrolyte lecture and another had the best CRRT session, everybody should benefit.

Over the years as we executed the first year fellow curriculum, we realized there was another curriculum that almost nobody was teaching.

A Bootcamp for Graduating Fellows. All the things you don’t realize you need until graduation is staring you in the face.

How to negotiate your first contract

What does a dialysis medical director actually do

How does value-based care affect your practice

How do you think about personal finance after years of training

Those aren’t board questions, but they’re the questions new nephrologists want to know but are afraid to ask. And let’s face it, the academic nephrologists that teach them nephrology would be poorly equiped to answer these questions.

So this year Bootcamp added a second day. The first day is aimed at second-year fellows who are getting ready to launch their careers. The second day is for incoming first-year fellows (and nephrocurious residents) who are just getting started.

The other surprise was how much the meeting benefited from being in person.

Sure, we could record the lectures (and we do). But the best parts happen during coffee breaks, lunch, and the conversations after the sessions end. Fellows meet faculty from other institutions. Faculty compare notes on how they teach. People swap slides, cases, and ideas. You can’t recreate that online.

This was our third Bootcamp, and our biggest yet.

I’m already thinking about next year. We’ll continue refining the curriculum, recording more sessions, and finding ways to share what we build with programs that can’t make it to Ann Arbor.

The original idea hasn’t changed.

Fellow education is too important for every program to go alone.

Kidney COn is a great conference and one of the highlights are the nights.

Thursday night is the faculty dinner. This year it included the students who got scholarships from KIDNEYcon to attend.

This year we went for dinner at Southern Tail Brewing and had endless appetizers followed by a pretty decent facsimile of a Chicago Dog

Then we played Corn Hole or as the Arkansas natives called it, Bags.

Roger and I came from behind for the win over Koyal and JC

The Friday evening plenary started with an extended cocktail hour followed by a blend of education and heart. There was a lively Nephrology Jeopardy competition MC’d by Bill Whittier of Rush (and the ASN Board Review). The KIDNEYcon put together a great Jeopardy Board that included these Answer and Question pairs for a category dedicated to nephrology authors.

This was followed by a career reflection from Roger Rodby. His talk, half physiology (exercise-associated hyponatremia), half memoir, highlighted the role of mentorship, serendipity, curiosity, and lawlessness in shaping an impactful career in nephrology.

The last night of KIDNEYcon is the best. It is the night John and Mary Arthur open their home for the famous Martini, Meatballs, and Mentorship party. It is always a great cap to a warm and inviting nephrology meeting.

And then when the night is over, there is always one more at the hotel bar.

Norris Jackson is something of a hero around Detroit. He was the first security guard at the Detroit Symphony Orchestra and remained their security guard for 34 years until he retired in 2023.

https://www.cbsnews.com/detroit/news/beloved-detroit-symphony-orchestra-security-guard-retires-after-34-years-of-service

Retirement didn’t slow him down. His Motor City Rams just won the Player Football Association Midwest Championship against their crosstown rivals, the Detroit Ravens.

But that wasn’t the Norris Jackson I knew. I knew the patient with severe hypertension and advanced chronic kidney disease.

I met Norris in 2019 when he was admitted with a blood pressure of 204/102, a creatinine of 5 mg/dL, and a potassium of 6.2. His prior creatinine from a year earlier was 3.4 (his whole story hinges on this earlier creatinine). We diagnosed him with hypertensive emergency, with AKI as the end-organ damage. We stabilized his blood pressure, but the kidney function never improved. Using the 2009 CKD-Epi race based formula his GFR was 13 ml/min. Using the current race-free 2021 formula it was 12 ml/min.

It is always hard to start the CKD journey in Stage 5. We were going to have speed run the advanced CKD playbook. He followed up in my clinic a couple of weeks after discharge and his blood pressure was improved to the 160s systolic. We referred him to transplant and cardiology for transplant clearance.

He returned a month later and his blood pressure was in the 140s. Cardiology started their work-up by scheduling an echo and a stress test with nuclear imaging. We increased his torsemide to 100 mg daily.

By June we had his blood pressure down to 127/60 on torsemide 100, chlorthalidone 25, nifedipine 60, and hydralazine 50 bid.

His transplant quest had the usual problems of collecting reports from other hospitals systems, dealing with a positive TB skin test, and completing multiple tests and imaging requests.

We also started him on epoetin and regular iron infusions. It was a lot but Norris did it all. Then a year after I met him, COVID hit and if navigating the health system seemed difficult before hand, it became near impossible during the pandemic. There will never be a full accounting how much damage was done by interrupting and delaying the care of everyone in the medical system.

Jackson wasn’t fully listed for a kidney transplant until November of 2022.

While he was navigating the world of transplant we were simultaneously preparing him for dialysis. Every visit we did modality education and discussion. He decided on home therapy and we referred to our general surgeon, Dr Meguid, who has really taken a deep interest in PD and become a valued partner. In June of 2021, Norris received a PD catheter and he began dialysis the following September. We did the whole process as an outpatient.

So by UNOS rules, even though he wasn’t listed until 2022, his wait time was backdated to the start of dialysis, September 2021.

But something else happened in 2021. The nephrology community abandoned race in the calculation of eGFR. While this change helped Black patients going forward, it did nothing for people whose transplant evaluations had already been delayed by the old race-adjusted equation. So in January 2023, the Organ Procurement and Transplantation Network (OPTN) required transplant programs to review Black kidney transplant candidates and determine whether earlier creatinine values would have made them eligible for listing using the race-free equation.

Norris had exactly that creatinine from a year before he presented to my service. This creatinine of 3.4 gave him a GFR of 21 by the 2009 formula, but with the 2021 race-free eGFR it was 19 ml/min. Low enough for him to be eligible for transplant listing.

With the stroke of a pen, his wait time moved from late 2021 to early 2018. An additional three years. This was enough to thrust him near the top of the waitlist.

In September of 2024, Norris got his kidney and three days later he was home with a falling creatinine, feeling better than he had in years.

A lot of people worked hard to remove race from eGFR equations. It can feel like an abstract policy debate.

But sometimes it looks like this.

Note: This was one of my first posts on Roon.com If you are an American physician who likes to chat about medicine, you should sign up.

There are so few prospective randomized trials in hyponatremia that we cherish every one, even the quirky, underpowered ones. So let’s take our hats off and salute the Royal Thai Air Force for putting the DDAVP clamp under the RCT microscope.

Safety and efficacy of proactive versus reactive administration of desmopressin in severe symptomatic hyponatremia: a randomized controlled trial

https://www.nature.com/articles/s41598-024-57657-z

The trial is underpowered: their power calculation called for 66 patients, but they enrolled 49. Still, I love that they had the courage to test something many of us, including me, have accepted without prospective data. Given the scarcity of trials, this is very much the hyponatremia way.

The investigators randomized patients to a proactive strategy (a DDAVP clamp) versus reactive DDAVP.

And they had the courage to enroll people who were actually sick. Many hyponatremia trials enroll patients with relatively mild disease, but this cohort had an average sodium of 115 mEq/L, which is impressive. Just look at the High risk of osmotic demyelination syndrome and the Clinical presentation in this population. These are the bad actors in hyponatremia—the patients who actually need randomized trials.

The HIT trial would have benefited from enrolling a cohort like this.

The results though, were similar to the HIT. There was no difference in overcorrection, 24-hour sodium change, or length of stay. The sodium rose more in the proactive group at 48 hours, but given multiple comparisons in a small trial, that signal should be ignored.

The overcorrection rate was 16.7% in the proactive group vs 28% in the reactive group (P = 0.54).

That’s remarkably consistent with MacMillan’s Toronto data, where about 18% of all comers overcorrected. It’s also far better than the 40% overcorrection rate reported by George Et al in Western Pennsylvania.

Interestingly, Pakchotanon enrolled patients with Na <125 mEq/L, neatly splitting the difference between the <130 threshold used by MacMillan and the <120 used by George.

Honestly, I expected better sodium control than they achieved. My suspicion is that the 3% saline dosing protocol in this trial was relatively aggressive. I know I tend to be more conservative with hypertonic saline when managing severe hyponatremia.

Still, this study is welcome. It’s part of a growing trend in hyponatremia research toward prospective data rather than retrospective dogma.

And in hyponatremia, every randomized trial moves the field forward.

One of the challenges in the assessment of hyponatremia is we rarely get to assess patients fresh from the community. By the time the nephrologist is called patients have usually received a liter (or two) of 0.9% normal saline. I call these sneaky IV fluids because they often don’t even get documented. Somewhere between triage and the nephrology consult a liter of saline appears—no order, no note, just a bag that quietly happened.

So when the emergency department, or ward, gets around to collecting the urine for biochemical analysis, the urine has been contaminated. Can we trust those altered specimens? Can they reliably distinguish hypovolemic hyponatremia from SIAD?

Well Dr Chienwichai of Thailand designed a study to answer that very question and he did it prospectively with a well designed rigorous trial.

He took patients with a sodium less than 130 and eliminated anyone with fluid overload. He also eliminated people where the urine sodium cannot be trusted:

• Nobody who received fluids before he got a chance to assess urine and blood
• Nobody on diuretics
• Nobody with adrenal insufficiency
• Nobody with metabolic alkalosis or even just a serum bicarbonate over 30
• Nobody with an eGFR less than 60

And then after checking basic urine and serum biochemical profiles ran everyone through protocolized fluid resuscitation. If the serum sodium failed to rise or fell with saline, patients were categorized as SIAD.If the sodium corrected with fluids and remained corrected when fluids were stopped, they were categorized as hypovolemic hyponatremia. The protocol looked something like this:

So at the end of the protocol they had the cohort convincingly divided into SIAD or hypovolemic hyponatremia. And they had a record of:

• Their initial urine and serum biochemical profile, Time Zero
• Their urine and biochemical profile after 500 ml, Time One
• After one liter, Time Two
• After two liters, Time Three
• and after four liters, Time Four

And so now we can see if that sneaky liter of saline ruins the biochemical tests. And you know what? It actually does the opposite. It makes the difference between SIAD and hypovolemic hyponatremia more stark. Here is table two with the averages:

So it is pretty clear that the numbers don’t meaningfully change with saline infusion but what I’ve been trying to explain in my recent lectures the urine sodium isn’t the window to truth we sometimes expect it to be. This question can be seen in Figure 2:

Though the numbers don’t lose information (and actually gain separation) with fluid resuscitation, there remains tremendous overlap with the urine Na between hypovolemic hyponatremia and SIAD. This is beautifully shown with the ROC curves the authors provide:

While an AUC of 0.75 is quite a bit better than 0.61, it’s helpful, but far from diagnostic.

One of the findings of the study was that patients who ultimately are diagnosed with SIAD often had an initial positive response to saline infusions:

However, our study showed that urine osmolality also decreases in SIAD from before saline infusion to after saline infusion. We hypothesize that this may be due to the vasopressin escape mechanism, a defense response characterized by increased urine volume and decreased urine osmolality. Furthermore, concurrent hypovolemia in hospitalized patients may have contributed to the observed decrease in urine osmolality. Although a decrease in the serum sodium level after saline administration is used as supportive criteria for diagnosis of SIAD25, approximately 30% of patients with SIAD have been reported to experience an increase in serum sodium by 5 mmol/L after receiving 2 L of 0.9% saline. In our study, this was observed in 38% of patients with SIAD.

I’ve seen this many times. SIAD can give a head fake by initially responding to saline. The sodium rises, everyone relaxes, and we think we were just treating volume depletion—only for SIAD to reappear a day later. It was satisfying to see that this was a real phenomenon rather than an undocumented spook.

The last finding I want to highlight is uric acid. I’m a fan of serum uric acid, and some of my fellows have been waving the flag for fractional excretion of uric acid (FEUA). However, in the baseline labs there was no significant difference in serum uric acid between hypovolemic hyponatremia and SIAD. FEUA did show a difference, but there was still substantial overlap. You can see this in how high the 75th percentile of FEUA was in hypovolemic hyponatremia, 13.9, even though the commonly cited cutoff for SIAD is only 12%.

The take home message:

Don’t panic if the urine studies were drawn after a liter of saline. The saline doesn’t destroy the diagnostic signal. If anything, it may accentuate the physiologic differences between hypovolemia and SIAD. And most importantly: urine sodium is helpful, but it is never the whole story.

The Hyponatremia Interventional Trial (HIT) has been published in NEJM Evidence. The results were unveiled at the Late-Breaking and High-Impact Clinical Trials session at ASN Kidney Week 2024 in San Diego, and we’ve been waiting 15 months for the manuscript to drop. Last week, it finally did.

https://evidence.nejm.org/doi/full/10.1056/EVIDoa2500086

The NEJM Evidence Journal has now published three of the most important articles on hyponatremia in the last few years. First with MacMillan Et al. that showed that CPM was quite a bit rarer than previously thought. Then the NEJM Evidence struck again with Seethapathy Et al. who showed that patients with hyponatremia who corrected at the slowest rate had higher hospital mortality. And now the HIT.

The premise of the HIT was straightforward and compelling. Hyponatremia has long been associated with worse outcomes: higher mortality, more rehospitalizations, longer lengths of stay. But association is not causation. So the investigators asked:

What if we deliberately and systematically corrected hyponatremia? If we do a better job at improving hyponatremia in one arm of a randomized trial and outcomes improve in that arm, that would go a long way to prove that hyponatremia is directly responsible for those adverse outcomes?

To test this, they randomized 2,100 patients to either standard care or a strategy of multifaceted targeted correction of hyponatremia. I would describe the intervention if it was describable but what it essentially comes down to is “consult nephro and have them run the hyponatremia playbook as described in Verbalis’ 2013 US guidelines and Spasovski’s 2014 European guidelines.” Take a look at figure 1 in the supplement. Give yourself more than a few minutes…it’s a lot.

The intervention worked, but only modestly.

The mean increase in serum sodium during the treatment period was:
• 10.0 mmol/L (±5.6) in the intervention group
• 8.7 mmol/L (±5.6) in the control group

Normal sodium levels (135–145 mmol/L) were achieved in:
• 60% of the intervention arm
• 46% of the control arm

Editorial side bar: It is depressing that the most current guidelines have not been updated for over a decade and fail to correct the sodium in 40% of patients…What are we doing here?

Back to the study: So yes, sodium moved more with the protocolized care. But the difference was modest.

And clinically? The primary outcome, death or rehospitalization at 30 days, occurred in:
• 20% of the intervention group
• 22% of the control group
• P = 0.45


Little separation, no signal. The authors essentially randomized patients to either standard care or a protocol that essentially looked a lot like standard of care. We should not be surprised that there was not separation. If we want to get separation it is time to move on from being afraid of rapid correction in low risk patients and crack the whip.

After decades of observational data linking hyponatremia to poor outcomes, HIT delivers a sobering message: correcting the number does not correct the prognosis. Or it is just a type 2 error due to the lack of separation between groups.